PROJECT SUMMARY The guidelines for initiation of colorectal cancer (CRC) screening are currently based on two risk factors: attained age and family history of CRC. Using the principles of precision medicine, we will individually tailor CRC screening recommendations based on the enormous knowledge we now have on genetic and non- genetic factors that predict risk for this disease. This strategy will reduce under and over-utilization of CRC screening, because individual risks vary substantially in the population, and over 80% of all CRC cases occur in those without a positive family history. In Aim 1 we will develop predictive models for CRC, based on genetic (~30M common and rare genotyped or imputed genetic variants) and clinico-epidemiologic variables (over 70 harmonized characteristics), derived from over 40,000 colorectal tumor cases and 46,000 controls. We will identify key predictors, and derive efficient personalized risk-prediction models for early detection of more treatable CRCs as well for CRC prevention, through the identification of advanced colorectal adenomas. In Aim 2 we will calibrate and validate these models in two prospective cohorts of >120,000 participants, including 40,000 minority members, diversity representing racially/ethnically and socioeconomically. These two cohorts (Research Program on Genes, Environment and Health and the Women's Health Initiative Minority cohort) contain genome-wide genotype array and comprehensive risk factor data coupled to data on screening and outcome data, allowing us to test the model's ability to predict risk for CRC and advanced colorectal adenoma across a broadly defined community-based population, and to personalize decision on starting age of screening based on individually specific genetic and environmental risk factors. In Aim 3 we will estimate the population benefit of our risk-stratified screening strategy, based on our risk prediction methods, compared with the current screening recommendations. This comparison will employ the well-tested decision model currently used to inform the United States Preventive Services Task Force CRC screening guidelines. Accomplishing these three aims, our research has the potential to accelerate the translation of a large amount of genetic and epidemiologic research to patient care by predicting advanced adenoma risk, for cancer prevention, and predicting cancer risk, for early cancer detection. Genetic testing is becoming part of routine care and genetic data will increasingly become part of an individual's medical record. Using genetic and non-genetic risk-factor information in clinical and preventive settings is a critical step towards developing precision medicine. Our models will provide recommendations for individually tailored CRC screening and interventions and, because they are personalized, may also increase adherence, maximize the appropriate use of invasive technologies, and guide important next steps towards public health policy development and clinical translation.